Study of genome packaging using single molecule manipulation and image method

My Ph.D research focuses mainly on the understanding of how these proteins perform their functions and on the study of DNA-protein interactional process by using magnetic tweezes and Ato

October, 2009

i

Acknowledgement

The works described in this thesis was carried out in the Biophysics & Single-molecule manipulation Lab, National University of Singapore (NUS), from August 2005 to July 2009, and was supported by research scholarship from the Physics department of NUS

I would like to thank Dr Yan Jie, my supervisor, for all his guide, help, support and encouragement when I was in his group for the past 4 years Without these, I would not have made so many achievements and it is an interesting and enriching experience for doing research and study in Biophysics & Single-molecule manipulation Lab

I am grateful to Dr Chen Hu, Dr Fu Hongxia, Dr Fu Wenbo, Law Dingying and all my group members for their help and suggestion during the period I am also grateful to my collaborator, Prof Peter Dröge, Prof Leong-Hew Choy, Prof Linda Kenney, and Dr Wu Jinlu for their excellent works and discussions

ii

Table of Content

Acknowledgement i

Table of Figures iv

Summary viii

Chapter 1 Introduction: Architectural Protein in Prokaryotes and Eukaryotes 1 1.1 References 14

Chapter 2 The techniques: Atomic Force Microscopy (AFM), Electrophoretic Mobility Shift Assay (EMSA), Transverse Magnetic tweezers 16

2.1 Atomic Force Microscopy 16

2.2 Mica surface modification 19

2.3 Magnetic Tweezers 22

2.4 Electrophoretic Mobility Shift Assay 25

2.5 References 26

Chapter 3 AFM study of scIHF-induced DNA bending 27

3.1 Introduction of IHF 27

3.2 Methods 31

Procedure for APTES functionalization 31

Procedure for Glutaraldehyde functionalization 31

AFM imaging of DNA-protein complexes 31

3.3 Results 32

3.4 Discussion 39

3.5 References 41

Chapter 4 Single DNA study of VP15-DNA interaction 43

4.1 Introduction 43

4.2 Methods 44

Electrophoretic mobility shift assay (EMSA) 44

Magnetic-tweezer Manipulation of VP15-DNA complex 45

4.3 Results 46

EMSA experiment confirmed that VP15 is a DNA-binding protein and it can package DNA cooperatively when the protein concentration exceeds a threshold value 46

Magnetic tweezer (MT) experiments revealed that VP15 could compact DNA against certain forces when the protein concentration was larger than a threshold value 49 AFM experiments revealed that VP15 packages DNA by making synergies

iii

between remote DNA sites 52

4.4 Discussion 53

4.5 References 57

Chapter 5 Single DNA study of H-NS-DNA interaction 59

5.1 Introduction 59

5.2 Methods 61

Magnetic-tweezer Manipulation of H-NS-DNA complex 62

Atomic Force Microscope imaging 62

5.3 Results 63

Ionic strength and magnesium ion alter the mode of H-NS binding to DNA 63

Magnesium acts as a switch between stiffening and bridging 65

Stiffening results from cooperative H-NS polymerization along DNA 68

During folding (bridging), large DNA hairpin structures form 71

5.4 Discussion 73

5.5 Supplementary Data 76

5.6 References 77

Chapter 6 Conclusion 80

List of publications 85

iv

Table of Figures Fig 1.1 [1] Cellular localization of the genome in cells from different kingdoms

of lives 3 Fig 1.2 (Picture is copied from Karolin Luger’s paper [8]) nucleosome core particle: ribbon traces for the 146-bp DNA phosphodiester backbones (brown and turquoise) and eight histone protein main chains (blue: H3; green: H4; yellow: H2A; red: H2B) The views are down the DNA superhelix axis for the left particle and perpendicular to it for the right particle For both particles, the pseudo-twofold axis is aligned vertically with the DNA centre at the top 6 

Fig 1.3 (Picture is copied from Karolin Luger’s paper [8]) The central base pair

through which the dyad passes is above the SHL0 label, (SHL, superhelix axis location) Each SHL label represents one further DNA double helix turn from SHL0 The complete histone proteins primarily associated with the 73-bp superhelix half are shown (interparticle tail regions are not shown) The two copies of each histone pair are distinguished as unprimed and primed copies, where the histone of the unprimed copy is primarily associated with the 73-bp DNA half and the primed copy with the 72-bp half The 4-helix bundles are labeled as H3’ H3 and H2B H4; histone-fold extensions of H3 and H2B are labeled as αN and αC, respectively; the interface between the H2A docking domain and the H4 C terminus as b; and

v

N- and C- terminal tail regions as N or C 7 

Fig 1.4 (Pictures are copied from Kerren K Swinger et al’s paper [11]) HU+DNA and IHF+DNA cocrystal structures 9 

Fig 2.1 Schematic diagram of an Atomic force microscope 18 

Fig 2.2 Pictures of AFM in our lab 19 

Fig 2.3 Schematic histogram showing the modified mica surfaces 21 

Fig 2.4 Schematic histogram of magnetic tweezers system 23 

Fig 2.5 Picture of magnetic tweezers system, including microscope and micro-manipulator 23 

Fig 2.6 Picture of flow channel and controlled magnet The glass with a 200µL tube on its left side is the channel inside which the DNA is attached (in the right part of the picture) The force is controlled by changing the distance between the magnet (the black bricks) and the channel 24 

Fig 3.1 Structure of IHF protein (Picture is copied from Phoebe A Rice et al [2]) 28 

Fig 3.2 Structure of single-chain IHF (scIHF) [10] 30 

Fig 3.3 AFM images of attL DNA on mica surface 35 

Fig 3.4 Zoom-in images of wild-type IHF induced DNA bending (The bright dot in the 2/3 part of DNA indicates a wild-type IHF in the expected location) 35 

Fig 3.5 Zoom-in images of scIHF2 induced DNA bending (The bright dot in the 2/3 part of DNA indicates a scIHF in the expected location) 36 

vi

Fig 3.6 Zoom-in images of scIHF2-K45αE induced DNA bending (The bright dot in the 2/3 part of DNA indicates a scIHF2-K45αE in the expected location) 36 Fig 3.7 Histogram of bending angle distribution 37 Fig 3.8 Histogram of bending angle distribution of Mg2+ dependence 37 Fig 3.9 Zoom – in image of scIHF2-K45αE induced DNA bending in 20 nM

Mg2+ solution condition (the bright dot in the 2/3 part of DNA indicates a scIHF2-K45αE in the expected location) 38 Fig 3.10 Zoom – in image of scIHF2-K45αE induced DNA bending in 200 nM

Mg2+ solution condition (the bright dot in the 2/3 part of DNA indicates a scIHF2-K45αE in the expected location) 38 Fig 4.1 Electrophoretic mobility shift assay (EMSA) 48 Fig 4.2 DNA folding dynamics under different forces and different VP15 concentrations 51 Fig 4.3 DNA unfolding dynamics under different forces and different VP15 concentrations 52 Fig 4.4 AFM images of linear phix174 DNA with and without VP15 (The height scale bar ranges from 0 – 2nm for Fig 4.4 a - d and f, and from 3 – 8nm for e) 53 Fig 5.1 Magnesium dependent binding modes of H-NS 66 Fig 5.2 H-NS interconverts between bridging and stiffening modes without being released from DNA 70 

vii

Fig 5.3 Imaging of DNA–H-NS complexes in the absence of or with low MgCl2concentration using Atomic Force Microscopy 71 Fig 5.4 Imaging of DNA–H-NS complexes in the bridging binding mode 72 Fig 5.5 Calcium substitutes for magnesium in stimulating the bridging/polymerization switch 76 Fig 5.6 Increasing the H-NS concentration dramatically reduces the DNA folding kinetics 76 

viii

Summary

The interaction between DNA and protein is of intense interest in biophysical research, especially the binding energy, DNA folding force, DNA elasticity and DNA-protein complex topography These are important in genomic compaction and function for all organisms

My Ph.D research focuses mainly on the understanding of how these proteins perform their functions and on the study of DNA-protein interactional process by using magnetic tweezes and Atomic Force microscopy (AFM) Magnetic tweezers is widely used in single DNA manipulation experiment and to study the dynamical process of DNA-protein interaction The static information, such as topography, of DNA-protein complexes can give the most direct evidence to assumptions which are derived from single DNA manipulation experiments The AFM is used to give structural details of DNA-protein complexes at the nano scale

In this thesis, I will describe 3 kinds of proteins that have been studied in my lab: Integration Host Factor (IHF), VP15 from White Spot Syndrome Virus (WSSV) and Histone-like Nucleoid Structural Protein (H-NS) All of them are DNA binding protein and have large influence on the DNA topography Our main interest is placed

on the topography of DNA-protein complexes, critical folding force and protein function under different ionic condition

VP15 shows the strongest DNA compacting ability among the 3 kinds of proteins with a critical folding force up to 5 pN However, IHF and H-NS are more interesting

ix

than VP15 in that their functions are ionic concentration dependent The bending ability of scIHF, presented by bending angle distribution of DNA-IHF structure, depends on Mg2+ concentration The most notable protein, H-NS, shows two switchable functioning modes according to whether Mg2+ or Ca2+ concentration is above certain value and two distinctly different DNA-H-NS structures are found using AFM

is expected to require more space than a flexible one One of the major challenges faced by many cells is how to effectively reduce the volume of its genome by several orders of magnitude while still retaining exactly all its genetic functionality and

effectiveness For example, most of the Escherichia coli cells are about 2 µm long and

0.5µm wide, but their chromosomal DNA molecules have a contour length of approximately 2 mm In absence of restriction, such a long DNA molecule would develop into a random coil whose volume is approximately 200 μm3 However, the

volume of an E.coli nucleoid is only around 0.5 μm3, around 1/400 of the unconstrained DNA size (Fig 1.1, copy from Martijn S Luijsterburg et al [1]) Therefore, there must be some mechanisms to operate to compact the chromosomal

2

DNA sufficiently into the cell

Macromolecular crowding is one of the mechanisms that are employed for compaction of DNA In cells, large amount of RNAs and proteins are produced from transcription of genomic DNA and translation of mRNAs, respectively The crowding condition caused by the high concentration of these macromolecules generates strong depletion/attraction forces [2, 3] Depletion force is one kind of entropic forces which arises when there are differently-sized particles in the solution The interactions among the excluded volumes of larger particles tend to yield large space between adjacent particles In turn, these increases of volume accessible to smaller particles result in strong attractive forces that can cause significant conformation changes of all particles [4] The concentration of RNAs and proteins in nucleoids and nuclei are within the range where depletion/attraction forces can occur and thus help to largely compact DNA It is highly possible that these crowded macromolecules in the cell induce considerable self-association of DNA and contributes to genomic organization Although depletion/attraction forces may contribute to the association of architectural proteins which has important effects on genomic folding, the resulting genome crowding results in only non-specific compaction Thus, the role it plays in gene regulating function may not be as important as it does in genomic compaction

3

Fig 1.1 [1] Cellular localization of the genome in cells from different kingdoms of lives

A) Microscopic image of a living human fibroblast (phase contrast) showing the nucleus by expression of a nuclear YFP-tagged DNA repair protein (DDB2)

B) Microscopic image of a living archaeal cell in late exponential phase of growth, showing

the nucleoid by staining with DAPI and Microscopic image of a living bacteria cell (E.coli) in

exponential phase of growth, showing the nucleoid by expression of GFP-tagged H-NS protein

To perform biological function such as transcription, translation, repression and derepression, architectural chromosomal proteins are necessary and most of them are small (~10 kDa), basic and have specific functions after binding to DNA According

to their effects on DNA, these proteins can be roughly divided into 3 classes: DNA wrapper, DNA bender and DNA bridger This classification is based mostly on the topographies of DNA-protein complexes, and it sheds light on understanding the fundamental role that those DNA architectural proteins played in the organization and

4

regulation of the genome Moreover, these architectural proteins show, to certain degree, conservative functionalities Therefore, the proteins from one kind of organism can perform similar function as proteins from another organism DNA compaction results generally from two modes: one is wrapping by histone/HMF protein in eukaryotes and some archaea [5], and the other is bending by HU/Sul7/Cren7/MC1 in bacteria and prokaryotes [6]

All kinds of Organisms have evolved their specific mechanisms to organize their genome and to compact it into different structures, like nucleoid in the prokaryotic cell and chromosomes in the eukaryotic cell

However, the binding affinity of the architectural proteins to DNA can be influenced by many factors such as ionic concentration, pH value of the surrounding environment, temperature and also the native structure of DNA For example, supercoiling DNA has high protein binding affinity, either by affecting the local DNA effective concentration through plectonemic formation which favors DNA bridging or

by reducing the free energy that is required to bend or wrap the DNA [1]

Because of the static repulsive force between DNA base pairs, there is a limited number of ways in which DNA structure can be regulated By inducing either bending

or bundling formation on DNA, architectural proteins, while reducing the effective volume, introduce functional regulation on genome A good example is the mentioned

“Histone-like nucleoid structural protein” (H-NS) whose binding can occur in the specific promoter region, thus preventing the access of RNA polymerase or other proteins to perform their tasks such as transcription and DNA cutting [7]

5

In eukaryotes, the compaction and regulation is mainly carried out by one kind of major proteins called histone proteins, which have histone-fold of 3-hydrophobic α-helices The histone proteins interact with DNA by inserting their arginine residues into the minor groove every helical turn on DNA [8] Generally, there exit 4 kinds of core histones: H2A, H2B, H3 and H4 They function by composing a “histone octamer” rather than work alone First, two H3-H4 connect together through their histone-fold, and then, two H2A-H2B associate with this “core” to form the octamer

~146 bp (basepair) of 200bp DNA wrap around the histone complex forming the nucleosome, which is widely known as a repressor to DNA-transacting processes such

as transcription (Fig 1.2 and Fig 1.3) It has been reported that the DNA is able to transiently detach from the octamer surface for 40-50ms, allowing proteins to access the previously warped DNA This may be one of the forms in which the transcription starts

The major DNA compaction mode in eukaryotes is histone induced DNA wrapping However, when it comes to bacteria and prokaryotes, the dominant mode of organizing DNA is bending and bridging because these organisms lack histone protein

In virus and prokaryotes, there is another type of proteins which is referred to as nucleoid-associated proteins (NAPs), because they also help DNA compaction by binding to DNA but lack structural resemblance with histones

6

Fig 1.2 (Picture is copied from Karolin Luger’s paper [8]) nucleosome core particle: ribbon traces for the 146-bp DNA phosphodiester backbones (brown and turquoise) and eight histone protein main chains (blue: H3; green: H4; yellow: H2A; red: H2B) The views are down the DNA superhelix axis for the left particle and perpendicular to it for the right particle For both particles, the pseudo-twofold axis is aligned vertically with the DNA centre at the top

7

Fig 1.3 (Picture is copied from Karolin Luger’s paper [8]) The central base pair through

which the dyad passes is above the SHL0 label, (SHL, superhelix axis location) Each SHL label represents one further DNA double helix turn from SHL0 The complete histone proteins primarily associated with the 73-bp superhelix half are shown (interparticle tail regions are not shown) The two copies of each histone pair are distinguished as unprimed and primed copies, where the histone of the unprimed copy is primarily associated with the 73-bp DNA half and the primed copy with the 72-bp half The 4-helix bundles are labeled as H3’ H3 and H2B H4; histone-fold extensions of H3 and H2B are labeled as αN and αC, respectively; the interface between the H2A docking domain and the H4 C terminus as b; and N- and C- terminal tail regions as N or C

Besides genomic organization and compaction, these proteins are involved in a broad range of DNA transacting processes such as replication, recombination,

8

transcription and DNA repair The proteins from bacteria and prokaryotes can generally be divided into two types according to their functions on DNA: DNA bender and DNA bridger The HU/IHF family is the best studied DNA bender [9, 10] This kind of protein is dimeric and composed by a compact core of α-helices and two flexible β-ribbon arms The two arms emanate from α-helices and insert into the minor groove of DNA to introduce a bending around the protein with an angle up to 160°。 HU shows a preferential binding to structural distortion on DNA sequence [11] (Fig 1.4), such as gaps, nicks Besides, HU can recognize pre-existing bending and help to stabilize it [12, 13] Moreover, the number of HU that binds to DNA can influence the bending angle and the extent to which DNA is compacted [14]

IHF has similar function on DNA as HU but it also shows DNA sequence specificity and non-specific binding [15] [16] However, the bend induced by HU is not as rigid as the one induced by IHF and it appears that HU can introduce a range of different bending angles similar to high mobility group (HMG) protein in eukaryotes Another widely studied DNA bender is Fis, which is shown to contribute significantly to nucleoid compaction both by first binding non-specifically to DNA and then bending or looping DNA [17] DNA loop formation is another function introduced by DNA bender and DNA bridger The DNA loops are one kind of nucleoid territories Their structures are dynamic and the boundaries are distributed in

a random manner [18] The different boundaries distribution along DNA can preserve the superhelicity of the genome but allow some of them to relax when proteins access DNA to perform their genomic function

B) The HU+DNA complex viewed from the top

C) Structure of IHF bound to the H’ site from phage I (Rice et al, 1996) The a subunit is white and the b subunit is pink and the intercalating prolines are yellow The DNA is blue except for the consensus sequence which is green

D) The sequence of the DNA substrate in TR3 with the three: T:T mismatches in pink and four unpaired “T”s in gray and green The “T”s are fipped out of the duplex and make crystal packing contacts in the structure, while the green “T”s remain stacked The yellow diamonds indicate sites of proline intercalation The light blue “C”s are partially disordered in TR3 structures

Another function of DNA bridging proteins is to cause dynamic cross linker between DNA strands which is either from one or several DNAs, forming a bundling

10

structure or large DNA – protein complexes The representative of DNA bridgers is

H-NS (Histone-liked nucleoid structural protein) which is found in E.coli and have

been wildly studied H-NS is a small abundant prokaryotic protein that organizes chromosomal DNA and plays an important role in gene silencing It serves as a negative regulator and represses the expression of many genes which are involved in bacterial adoption to environmental change [19] H-NS is present at 20,000 copies per cell and binds preferentially to A-T rich segments on DNA It is composed of 3 parts and each has different function The C-terminal domain (residues 90-136) binds DNA, while the N-terminus (residues 1-64) is involved in H-NS dimerization The two domains are connected via an unstructured linker which is comprised of residues 65-89 H-NS exists as a dimer and has the ability to self-associate, forming higher order oligomers With the two DNA binding sites, H-NS dimmer can interact with two DNA strands simultaneously [9] Recent research has shown that H-NS has no DNA sequence preference but does have high affinity to AT rich parts [20] By binding to the promoter region of many genes, H-NS can inhibit RNA polymerase and other proteins from accessing, thus interfering with transcription initiation In this situation, H-NS functions as a repressor, silencing selectively specific genes or regions of chromosome, and as coordinator acting in concert with other transcription factors Recent detailed research suggests that H-NS binds first to nucleating high-affinity sites separately, and then, dimmer – dimmer interaction leads to H-NS polymerization that results eventually in the formation of a supercoiled intertwined filament containing two DNA duplexes connected by protein bridges, constraining a DNA loop

11

[20] The bundling DNA filament covered with H-NS nucleation would silence an extensive region of genes and operons Seperate biophysical studies have revealed how H-NS dimers can bridge DNA and how the cooperative polymerization can happen and block transcription initiation[21]

When the network of DNA – H-NS complexes forms, DNA is sequestered from promoters which are regulated When the surrounding environment changes, such as a decrease in ionic concentration, lack in food supply or temperature variation, in adaption to the new condition, H-NS proteins should leave from or attach to specific promoter sites This leads to the hypothesis that there may be a structural reorganization of the nucleoid The factor is receptive to signals from changed environmental conditions Another possibility is that H-NS can act as an environmental sensor by changing its cooperative mode leading to new DNA - protein complexes (This effect has been showed by our group) According to our results, when Mg2+ is present, H-NS behaves as a DNA compactor, comparing to function as a DNA stiffener in absence of Mg2+, and this will be discussed in detail in a later chapter

To date, the mechanism and function of how nucleoid-associated proteins (NAPs) compact and regulate genomic expression is still unclear Besides, none of these protein functions alone in DNA compactions, which is totally different from the situation in eukaryotes where the histone protein contribute exclusively to DNA

compaction Although, in in vitro experiments all NAPs clearly exhibit the ability to condense DNA, while in in vivo experiments, the effect from individual protein alone

12

is limited For example, bacteria lacking one of the NAPs usually have subtle phenotypes, which suggests there is an overlap among the protein functions and the role of one protein can be compensated by another

Another important phenomenon is that the expression level of nucleoid-associated proteins depends largely on cell growth phase T A Azam et al analyzed 12 nucleoid – associated proteins at different cell growth phases [22] The result showed that during exponential growth, some proteins are absent and become abundant during the stationary growth, or vice versa This may be a possible explanation of how the cell can meet the requirement fro high levels of transcription and translation during growth and how to protect its genome when the stationary phase comes The different expression level at different cell growth phage gives hint

to model DNA structure according to the growth condition For example, cell increases the Fis concentration to stimulate the transcription of stable RNA operons during growth and expresses Drp to bind extensively on genome to stop transcription

at stationary phase In order to switch between compacting and relaxed states, some of the nucleiod – associated proteins have definite genome condensing capability while some of the proteins have dual function and can act either as compacting agent or as antagonists For example, Fis and IHF, which are DNA benders, can de-repress the impeditive effect of H-NS at specific promoters HU exhibits a similar ability using a different mechanism which is to compete for preferential binding sites However, more complex mechanisms may exist and require many proteins to work

cooperatively in in vivo condition since there are hundreds of proteins inside a cell

of architectural protein that the organisms have developed However, the number of options to reduce the volume of genome seems limited and they are used in all forms

of living organisms

In later chapters, organization and compaction of DNA by 3 kinds of proteins are studied using Atomic Force Microscope (AFM) and transverse magnetic tweezers The results indicate that the protein from virus has a simpler function compared with the proteins from prokaryotes IHF introduces different DNA bending angles according to variation of Mg2+ concentration and H-NS, being more complicated, has two binding modes which lead to two distinct DNA structure

14

1.1 References

[1] M S Luijsterburg, M F White, R v Driel, and R T Dame, Critical Reviews in

Biochemistry and Molecular Biology 43, 393 (2008)

[2] D Marenduzzo, K Finan, and P R Cook, Journal of cell biology 175, 681

(2006)

[3] R Hancock, Journal of structural biology 146, 281 (2004)

[4] P Nelson, W.H Freeman and Company, New York (2004)

[5] M S Luijsterburg, M C Nooma, G J L Wuite, and R T D a, Journal of

structural biology 156, 262 (2006)

[6] K K Swinger, and P A Rice, Current Opinion in Structural Biology 14, 28

(2004)

[7] C J Dorman, Nature reviews Microbiology 5, 157 (2007)

[8] K Luger, A W Mäder, R K Richmond, D F Sargent, and T J Richmond,

[15] P A Rice, S.-w Yang, K Mizuuchi, and H A Nash, Cell 87, 1295 (1996)

[16] B M J Ali, R Amit, I Braslavsky, A B Oppenheim, O Gileadi, and J Stavans,

Proceedings of the National Academy of Sciences 98 (2001)

[17] D Skoko, J Yan, R C Johnson, and J F Marko, Physical review letters 95,

Travers, Nucleic Acids Research 35, 6330 (2007)

[21] E Bouffartigues, M Buckle, C Badaut, A Travers, and S Rimsky, nature

structural & molecular biology 14, 441 (2007)

[22] T A Azam, A Iwata, A Nishimura, S Ueda, and A Ishihama, Journal of

Bacteriology 181, 6361 (1999)

16

Chapter 2 The techniques: Atomic Force Microscopy (AFM), Electrophoretic Mobility Shift Assay (EMSA), Transverse Magnetic tweezers

2.1 Atomic Force Microscopy

DNA, although very long, has a diameter of merely 2 nm which makes it impossible to be visualized by traditional microscopy To see nano-sized DNA-protein complexes, SPMs (scanning probe microscopy) are widely applied in biological research For instance, SEM (scanning electron microscope) and TEM (transmission electron microscope) are able to show clear images for minimum sample-size 10nm and many images of nucleus and nucleoid has been reported However, there are some drawbacks for these EM imaging techniques In order to acquire high resolution images, both the samples and substrates must be conductive Moreover, the sample should be fixed, i.e not moving freely on the surface These require the chemical modification of substrates, which may induce unknown artificial results on the DNA-protein complexes such as DNA condensation and protein repulsion Another problem is the surface extension force that stretches DNA when a sample is placed in the vacuum chamber

These limitations were addressed with the invention of atomic force microscopy (AFM) which has some special advantages when compared to normal Electron

17

Microscopy (EM) First, it does not require the sample to be conductive and is capable

of measurement of the topography of almost any kind of surfaces Second, AFM uses

a reflection laser from the probe that scans the sample by gently “touching” the surface This can effectively reduce the perturbation and damage of biological samples when observed with EM

The main components of a modern AFM are mainly a cantilever, a laser beam deflection system, a piezoelectric scanner nose, an electronic control unit and a computer that controls the whole system (Fig 2.1)

There are two primary scanning modes of AFM: contact mode and tapping mode (or AC mode), according to whether the probe contacts with the sample surface constantly or not

In contact mode, the cantilever probe is brought into physical contact with the scanned sample and tracks the topographic changes as the probe moves along the surface This causes the deflection in the cantilever to deflect which then change the position of the reflecting laser on a four quadrant photodetector The change is then calculated and converted to the morphological change by the electronic control unit and computer program

When the probe moves across the samples surface under the contact mode, the sharp tip and the lateral component can cause damage to soft or fragile samples, especially in the case of biological specimens In these cases, tapping mode is preferred Under tapping mode, the probe is not continuously in contact with sample surface like that in contact mode Instead, the probe oscillates during scanning, and a

18

force interaction between the probe and the sample causes a change in the resonant frequency and oscillation amplitude of the vibrating cantilever Either of them is then used to control the tracking of the probe over the surface This mode allows the same high sensitivity compared to contact mode without causing damage to the soft samples

In our experiment, tapping mode AFM was used to scan DNA-protein complexes

on differently modified mica surfaces (Fig 2.2)

Fig 2.1 Schematic diagram of an Atomic force microscope

19

Fig 2.2 Pictures of AFM in our lab

All the AFM images are acquired using tapping mode scanning The cantilevers are made of silicon with resonance frequency at 240 kHz to 300 kHz and spring constant at ~40N/m For stability and quality reasons, all the scans are performed under 1 Hz line rate and 512 x 512 resolution

2.2 Mica surface modification

AFM is proven to be an extremely useful instrument in biological research, especially in the study DNA-protein interaction Despite its advantages and the progress made, a major handicap stems from the unreliable nature of the deposition

20

process in which only a small amount of sample can be found on the mica surface DNA has a diameter of merely 2 nm and the proteins have normal sizes of 1~3nm, therefore, any surface roughness larger than 2 nm will have great influence on the scanning results, hindering useful information of conformational details Mica is a good candidate for DNA-protein complex imaging because of the smoothness of its surface (height fluctuation is less than 1 Å) However, the issue is that when contacting with water, the mica surface becomes negatively charged, thus DNA is repelled away from the surface because it is also negatively charges Any washing and drying process leave little DNA for imaging

To date, most of the surface modifications are based on the electrostatic attachment of bio-samples to an oxide surface A common method is to use bivalent ion, such as Mg2+ or Ca2+, to place positive charges on the mica surface serving as a bridge between the mica surface and DNA molecules (Fig 2.3 A) However, these cations, while helping to bind DNA to the negatively charged surface, can also condense DNA by neutralizing the intrinsic charges of DNA, resulting in unpredictably compacting effect Besides, bivalent ion effect on protein function is of great interest to scientists, so depletion of Mg2+ or Ca2+ is necessary and this prevents the application of saline fixation of DNA

Another modification places amine on the mica surface by reaction with aminopropyltriethoxysilane (APTES) and the amine group can hold DNA tightly without bivalent ions (Fig 2.3 B) Despite its excellent capability in adhering the DNA random coiled structure for AFM imaging, the APTES-modified mica has been

Fig 2.3 Schematic histogram showing the modified mica surfaces.

22

2.3 Magnetic Tweezers

SPM (Scanning probe microscopy) can provide details of DNA-protein structures with high resolution, but lacks the liability to reveal details about the dynamical process and force response which are of great interest in biophysics Magnetic tweezers is an instrument that, by using magnetic gradient field, exerts and measures the force on magnetic beads Its typical application is in micromanipulation of single DNA molecules In brief, the two ends of a DNA molecule (normally the 48.5kb -DNA) is first labeled with biotin- and digoxygenin-labeled oligonucleotides, respectively [2] One end of labeled DNAs is then bound to a 2.8-micron-diameter paramagnetic bead and the other end to an edge of a thin 0# cover glass By placing a magnet near the paramagnetic bead, a force stretching the tethering DNA molecules is applied on the bead, and the force is calculated from the measurement of the bead’s

Brownian motion transverse to the direction of the force using the equation: F/L =

kBT/(δX2) Here, L the extension of DNA; (δX2) represents an average over the square

of the bead transverse displacement; T is the temperature and the kB is Boltzmann’s constant In our experiment, the range of force is from 0.1pN to 20pN The force was applied on the focal plane of the objective, and the extension of DNA was determined

by measuring the distance between the bead and the edge of the cover glass in the force direction More detailed description of the experiment setup and reference will

23

be introduced in Chapter 4

Using magnetic tweezers, it is possible to directly “observe” the biological process such as DNA folding and DNA stiffening Development of the techniques for manipulation of single DNAs is of large interest to biological physicists and molecular biologists

Fig 2.4 Schematic histogram of magnetic tweezers system

Fig 2.5 Pictures of magnetic tweezers system, including microscope and micro-manipulator

24

Fig 2.6 Picture of flow channel and controlled magnet The glass with a 200µL tube on its left side is the channel inside which the DNA is attached (in the right part of the picture) The force is controlled by changing the distance between the magnet (the black bricks) and the channel

25

2.4 Electrophoretic Mobility Shift Assay

The Electrophoretic Mobility Shift Assay (EMSA), also referred to as gel retardation array or gel shift array, is a common technique used to characterize DNA-protein/RNA-protein interaction It can determine whether a protein is capable

of binding to a DNA/RNA and to cause structural changes of DNA/RNA such as bending, relaxing or cutting EMSA is based on the observation that DNA-protein/RNA-protein complexes migrate through a nondenaturing agarose gel more slowly than free DNA/RNA fragments In brief, the mixtures of DNA-protein/RNA-protein are loaded into agarose gel, and an external electric field

is applied to drive complexes moving along according to the charge of the DNA-protein/RNA-protein mixture The speed at which these molecules move in the gel is determined by their charges, sizes and their shapes A control lane usually contains unbound DNA/RNA Then, assuming that the protein binds to the DNA/RNA fragment, the lanes with DNA-protein/RNA-protein complexes will contain one or several different bands that represent the larger, less mobile complexes since the binding of protein can change not only charges, but also sizes and shapes of the complexes

The integration host factor (IHF) is a key DNA architectural protein in

Escherichia coli [1] IHF has two homologous subunits, the α- and the β-subunit and

it shows limited DNA sequence preference Both subunits are ~10 kDa and are ~30% identical in sequence The two subunits are intertwined to form a compact structure, from which two long β ribbon “arms” extend (Fig 3.1) [2]

28

Fig 3.1 Structure of IHF protein (Picture is copied from Phoebe A Rice et al [2])

a) The α - subunit and β - subunit are shown in white and pink, respectively Double helix DNA is represented in blue and the green part on it is the consensus sequence which interacts mainly with the arm of α – subunit and the body of β – subunit

b) The top view of IHF – DNA complex

c) Sequence and secondary structure of the two subunits

d) The crystal structure of IHF-DNA complexes (5 asymmetric units are shown)

Many works have been done on IHF-DNA snups, including crystal structure of IHF in complex with the phage λ H’ site, fluorescent resonance energy transfer (FRET) analyses and visualization of IHF-DNA complexes through atomic force microscopy [2-4] All of these revealed a strong protein-induced DNA bending (120-160°) The bending is mainly caused by the intercalation of one conserved proline residue from each subunit of heterodimer in to the DNA minor groove By electrostatic interaction

29

round the protein body, the DNA U-turn conformation is then stabilized [5, 6]

IHF is involved in the regulation of more than 100 genes in Gram-negative bacteria and it is an essential cofactor in phage λ site-specific recombination, where the protein serves an architectural role during the assembly of snups [7, 8] Phage attachment (att) site attP, composed of 240 bp, is one of the two recombination sequences in the integrative pathway and harbors three specific IHF-binding sites which must be occupied by IHF to form a functional snup, the so-called integrative intasome Moreover, negative DNA supercoiling of attP is necessary for intasome assembly [9] The intasome then captures the protein-free 21 bp attB to form a synaptic complex in which two successive rounds of DNA strand exchange are catalyzed by phage λ integrase (Int)

Our collaborator transferred the phage λ recombination system to mammalian cells and engineered a single-chain IHF, named scIHF2, which is functional in mammalian cells The scIHF2 differs from wild-type IHF in that almost the entire α-subunit is inserted into the β-subunit at position 39 using two short peptide linkers (Fig 3.2) [10] Their biochemical and functional assays confirmed that scIHF2 behaves like its heteromeric parent A variant of scIHF2, called scIHF2-K45αE was also identified, and it carries glutamate instead of lysine at position 45 of the α-subunit[10] Besides, one of the two linkers is shortened in scIHF2-K45αE This

new variant is found to be nearly inactive in promoting integrative recombination in

vitro, while remaining fully active as a co-factor for excisive recombination on

supercoiled DNA The protein also exhibits a defect in its function as an initiation

30

factor for pSC101 replication in vivo

Fig 3.2 Structure of single-chain IHF (scIHF) [10]

A) ScIHF2-H’structure The two linkers (labeled 1 and 2) used to connect the two subunits in scIHF2 are highlighted in cyan The DNA is depicted in purple

B) Zoomed-in image highlighting the two linkers and the respective residues that were chosen

to connect the two IHF subunits

In this research, the interesting phenotype is analyzed in detail by introducing the K45αE substitution into scIHF2 This leads to the identification of a novel, controllable modular mode of protein-induced DNA bending In addition, the results obtained with the phage λ site-specific recombination system provide valuable insight into possible dynamics of functional snup formation in general, and how this can be governed by an intricate interplay between DNA architectural proteins and external factors